Clustering criteria in multiobjective data clustering

Julia Handl; Joshua Knowles

doi:10.1007/978-3-642-32964-7_4

Clustering criteria in multiobjective data clustering

Julia Handl, Joshua Knowles

MSM Decision Sciences

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Abstract

We consider the choice of clustering criteria for use in multiobjective data clustering. We evaluate four different pairs of criteria, three employed in recent evolutionary algorithms for multiobjective clustering, and one from Delattre and Hansen's seminal exact bicriterion method. The criteria pairs are tested here within a single multiobjective evolutionary algorithm and representation scheme to isolate their effects from other considerations. Results on a range of data sets reveal significant performance differences, which can be understood in relation to certain types of challenging cluster structure, and the mathematical form of the criteria. A performance advantage is generally found for those methods that make limited use of cluster centroids and assess partitionings based on aggregate measures of the location of all data points. © 2012 Springer-Verlag.

Original language	English
Title of host publication	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)\|Lect. Notes Comput. Sci.
Place of Publication	Berlin/Heidelberg
Publisher	Springer Nature
Pages	32-41
Number of pages	9
Volume	7492
ISBN (Print)	9783642329630
DOIs	https://doi.org/10.1007/978-3-642-32964-7_4
Publication status	Published - 2012
Event	12th International Conference on Parallel Problem Solving from Nature, PPSN 2012 - Taormina, Italy Duration: 1 Sept 2012 → 5 Sept 2012

Conference

Conference	12th International Conference on Parallel Problem Solving from Nature, PPSN 2012
Country/Territory	Italy
City	Taormina
Period	1/09/12 → 5/09/12

Keywords

multiobjective clustering, bicriterion clustering, clustering objectives, multiple clustering objectives

Access to Document

10.1007/978-3-642-32964-7_4

Cite this

@inproceedings{426a48daa867499eb7b0c42b668c571c,

title = "Clustering criteria in multiobjective data clustering",

abstract = "We consider the choice of clustering criteria for use in multiobjective data clustering. We evaluate four different pairs of criteria, three employed in recent evolutionary algorithms for multiobjective clustering, and one from Delattre and Hansen's seminal exact bicriterion method. The criteria pairs are tested here within a single multiobjective evolutionary algorithm and representation scheme to isolate their effects from other considerations. Results on a range of data sets reveal significant performance differences, which can be understood in relation to certain types of challenging cluster structure, and the mathematical form of the criteria. A performance advantage is generally found for those methods that make limited use of cluster centroids and assess partitionings based on aggregate measures of the location of all data points. {\textcopyright} 2012 Springer-Verlag.",

keywords = "multiobjective clustering, bicriterion clustering, clustering objectives, multiple clustering objectives",

author = "Julia Handl and Joshua Knowles",

year = "2012",

doi = "10.1007/978-3-642-32964-7_4",

language = "English",

isbn = "9783642329630",

volume = "7492",

pages = "32--41",

booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)|Lect. Notes Comput. Sci.",

publisher = "Springer Nature",

address = "United States",

note = "12th International Conference on Parallel Problem Solving from Nature, PPSN 2012 ; Conference date: 01-09-2012 Through 05-09-2012",

}

Handl, J & Knowles, J 2012, Clustering criteria in multiobjective data clustering. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)|Lect. Notes Comput. Sci.. vol. 7492, Springer Nature, Berlin/Heidelberg, pp. 32-41, 12th International Conference on Parallel Problem Solving from Nature, PPSN 2012, Taormina, Italy, 1/09/12. https://doi.org/10.1007/978-3-642-32964-7_4

Clustering criteria in multiobjective data clustering. / Handl, Julia ; Knowles, Joshua.
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)|Lect. Notes Comput. Sci.. Vol. 7492 Berlin/Heidelberg: Springer Nature, 2012. p. 32-41.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Clustering criteria in multiobjective data clustering

AU - Handl, Julia

AU - Knowles, Joshua

PY - 2012

Y1 - 2012

N2 - We consider the choice of clustering criteria for use in multiobjective data clustering. We evaluate four different pairs of criteria, three employed in recent evolutionary algorithms for multiobjective clustering, and one from Delattre and Hansen's seminal exact bicriterion method. The criteria pairs are tested here within a single multiobjective evolutionary algorithm and representation scheme to isolate their effects from other considerations. Results on a range of data sets reveal significant performance differences, which can be understood in relation to certain types of challenging cluster structure, and the mathematical form of the criteria. A performance advantage is generally found for those methods that make limited use of cluster centroids and assess partitionings based on aggregate measures of the location of all data points. © 2012 Springer-Verlag.

AB - We consider the choice of clustering criteria for use in multiobjective data clustering. We evaluate four different pairs of criteria, three employed in recent evolutionary algorithms for multiobjective clustering, and one from Delattre and Hansen's seminal exact bicriterion method. The criteria pairs are tested here within a single multiobjective evolutionary algorithm and representation scheme to isolate their effects from other considerations. Results on a range of data sets reveal significant performance differences, which can be understood in relation to certain types of challenging cluster structure, and the mathematical form of the criteria. A performance advantage is generally found for those methods that make limited use of cluster centroids and assess partitionings based on aggregate measures of the location of all data points. © 2012 Springer-Verlag.

KW - multiobjective clustering, bicriterion clustering, clustering objectives, multiple clustering objectives

U2 - 10.1007/978-3-642-32964-7_4

DO - 10.1007/978-3-642-32964-7_4

M3 - Conference contribution

SN - 9783642329630

VL - 7492

SP - 32

EP - 41

BT - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)|Lect. Notes Comput. Sci.

PB - Springer Nature

CY - Berlin/Heidelberg

T2 - 12th International Conference on Parallel Problem Solving from Nature, PPSN 2012

Y2 - 1 September 2012 through 5 September 2012

ER -

Clustering criteria in multiobjective data clustering

Abstract

Conference

Keywords

Access to Document

Fingerprint

Cite this